Streamlining Customer Service with a Deep Learning-Powered Chatbot


Sakha’s client was from the investment industry, which is highly competitive and customer-centric. Investment firms are always looking for ways to improve customer experience and satisfaction while reducing costs. In such a scenario, the need for a chatbot solution arises. Sakha helped the client streamline customer service with a deep learning-powered chatbot.

The Challenge

The client faced several challenges, which warranted the need for a chatbot solution such as:

  • High volume of customer queries: Investment firms receive a high volume of customer queries related to onboarding, cash flow, stock portfolio, etc. This can be overwhelming for customer service agents to handle manually.
  • 24/7 availability: Customers may have queries outside of regular business hours, and it is not feasible to have customer service agents available at all times.
  • Cost-effective solution: Hiring and training more customer service agents to handle the high volume of queries can be expensive. Therefore, the client needed an efficient and cost-effective solution.


Sakha used a deep learning model to develop & deploy a chatbot-based solution to understand and automatically reply to customer queries. The model was trained on a large dataset of customer queries and responses to understand the context of the queries and provide accurate responses. As the chatbot interacts with more customers, the model learns and improves its responses over time. The chatbot can also handle complex queries by understanding the intent behind them and providing relevant responses. Some of the functions which the chatbot was programmed to handle included:

  • Onboarding: The chatbot can guide new customers through the onboarding process by asking questions, providing necessary information and collecting the responses.
  • Cash Flow Queries: Customers can ask queries related to their cash flow, such as account balance, transactions, and payments.
  • Stock Portfolio Queries: Customers can ask queries related to their stock portfolio, such as stock prices, performance, and investments.
  • Notifications: The chatbot can send notifications to customers regarding account updates, market trends, and investment opportunities.
  • News Feeds: The chatbot can provide relevant news and information related to the investment industry and the customer’s stock portfolio.
  • Reporting: The chatbot can generate reports related to the customer’s account and portfolio.
For developing the chatbot, Sakha leveraged on Rasa for the Chatbot Development Framework, Python, PyTorch for the Deep Learning Framework, NLTK for the Natural Language Processing Library, Django for the Backend Framework, PostgreSQL and AWS for the Cloud Platform. Benefits of the chatbot solution include:
  • 24/7 availability: The chatbot is available to customers 24/7, providing immediate responses to their queries.
  • Cost-effective solution: The chatbot can handle a high volume of queries at a low cost compared to hiring more customer service agents.
  • Improved customer experience: The chatbot can provide quick and accurate responses to customer queries, improving customer satisfaction and experience.
  • Integration: The chatbot was integrated with the investment firm’s app, customer software, Facebook page, and website, providing a seamless customer experience.